import streamlit as st
import tensorflow as tf
from tensorflow.keras.preprocessing import image
import numpy as np
import matplotlib.pyplot as plt
from PIL import Image
# Streamlit app
st.title("Deep Learning for Military Aircraft Recognition in Satellite Imagery")
# Load the model
model = tf.keras.models.load_model(
"MilSat224.keras", compile=False, safe_mode=False
)
# Create a mapping for class indices to class names
class_indices = {
0: 'A-10',
1: 'B-1',
2: 'B-2',
3: 'B-52',
4: 'Bareland',
5: 'C-130',
6: 'K/C-135',
7: 'C-17',
8: 'C-5',
9: 'E-3',
10: 'KC-10',
}
#################################################################
ellsworth = """
"""
macdill = """
"""
whiteman = """
"""
elmendorf = """
"""
minot = """
"""
ramstein = """
"""
mcguire = """
"""
#######################
eielson = """
"""
beale = """
"""
edwards = """
"""
vandenberg = """
"""
andrews = """
"""
cannon = """
"""
altus = """
"""
andersen = """
"""
col4 = st.columns(1)
with col4[0]:
st.subheader("Aircraft Types Included in Model Training:")
st.markdown("""
- A-10
- B-1
- B-2
- B-52
- C-130
- K/C-135
- K/C-10
- C-17
- C-5
- E-3
- Bare Terrain
""")
st.subheader("Detailed Workflow:")
st.write("1. Select an available base from the dropdown menu or navigate to a base using google maps")
st.write("2. Identify a suitable aircraft parked on the ramp")
st.write("3. Use a capture tool to screenshot a satellite image of selected singular aircraft")
st.write("4. Upload your captured image")
st.write("5. View model prediction of aircraft type")
selected_base = st.selectbox("Select a base to view satellite imagery:", ["Ellsworth AFB","MacDill AFB","Whiteman AFB",
"Elmendorf AFB","Minot AFB","Ramstein AB",
"McGuire AFB","Eielson AFB","Beale AFB",
"Edwards AFB","Vandenberg AFB","Andrews AFB",
"Cannon AFB","Altus AFB","Andersen (Guam)"])
# Display the selected map
if selected_base == "Ellsworth AFB":
st.markdown(ellsworth, unsafe_allow_html=True)
elif selected_base == "MacDill AFB":
st.markdown(macdill, unsafe_allow_html=True)
elif selected_base == "Whiteman AFB":
st.markdown(whiteman, unsafe_allow_html=True)
elif selected_base == "Elmendorf AFB":
st.markdown(elmendorf, unsafe_allow_html=True)
elif selected_base == "Minot AFB":
st.markdown(minot, unsafe_allow_html=True)
elif selected_base == "Ramstein AB":
st.markdown(ramstein, unsafe_allow_html=True)
elif selected_base == "McGuire AFB":
st.markdown(mcguire, unsafe_allow_html=True)
elif selected_base == "Eielson AFB":
st.markdown(eielson, unsafe_allow_html=True)
elif selected_base == "Beale AFB":
st.markdown(beale, unsafe_allow_html=True)
elif selected_base == "Edwards AFB":
st.markdown(edwards, unsafe_allow_html=True)
elif selected_base == "Vandenberg AFB":
st.markdown(vandenberg, unsafe_allow_html=True)
elif selected_base == "Andrews AFB":
st.markdown(andrews, unsafe_allow_html=True)
elif selected_base == "Cannon AFB":
st.markdown(cannon, unsafe_allow_html=True)
elif selected_base == "Altus AFB":
st.markdown(altus, unsafe_allow_html=True)
elif selected_base == "Andersen (Guam)":
st.markdown(andersen, unsafe_allow_html=True)
# Upload image
uploaded_file = st.file_uploader("Upload an image", type=["jpg", "jpeg", "png"])
if uploaded_file is not None:
# Open and display the uploaded image
img = Image.open(uploaded_file).convert("RGB")
st.image(img, caption='Uploaded Image', width=300)
# Preprocess the image
img = img.resize((224, 224)) # Resize to 224x224
img_array = image.img_to_array(img) # Convert to array
img_array = np.expand_dims(img_array, axis=0) # Expand dimensions
img_array /= 255.0 # Normalize if needed
# Make prediction
predictions = model.predict(img_array)
predicted_class_index = np.argmax(predictions, axis=1)[0]
predicted_class_name = class_indices.get(predicted_class_index, "Unknown Class")
# Display the result
st.metric("Predicted Aircraft:", predicted_class_name)